Publications

Journal and conference papers, preprints and theses. Bibliometrics on Google Scholar; everything as BibTeX or plain text.

2026

CHANI: Correlation-based Hawkes Aggregation of Neurons with bio-InspirationS. Jaffard, S. Vaiter, and P. Reynaud-BouretJournal of Machine Learning Research, 2026arXiv
Deep Equilibrium models for Poisson Imaging Inverse problems via Mirror DescentC. Daniele, S. Villa, S. Vaiter, and L. CalatroniSIAM Journal on Imaging Sciences, 2026DOI arXiv
Bilevel gradient methods and Morse parametric qualificationJ. Bolte, T. Lê, E. Pauwels, and S. VaiterMathematics of Operations Research, 2026DOI arXiv
Towards Understanding Steering StrengthM. Taimeskhanov, S. Vaiter, and D. GarreauInternational Conference on Machine Learning (ICML), 2026arXiv
Fairness-informed Pareto Optimization: An Efficient Bilevel FrameworkS. Tanji, S. Vaiter, and Y. LaguelPreprint, 2026arXiv
ZOBA: An Efficient Single-loop Zeroth-order Bilevel Optimization AlgorithmM. Rando, and S. VaiterPreprint, 2026arXiv
On the Hardness of Junking LLMsM. Rando, and S. VaiterPreprint, 2026arXiv
Proximal basin hopping: global optimization with guaranteesG. Lauga, C. Molinari, and S. VaiterPreprint, 2026arXiv
Characterizations of inexact proximal operatorsG. Lauga, and S. VaiterPreprint, 2026arXiv

2025

Risk Estimate under a Nonstationary Autoregressive Model for Data-Driven Reproduction Number EstimationB. Pascal, and S. VaiterSignal Processing, 2025DOI arXiv
Geometric and computational hardness of bilevel programmingJ. Bolte, T. Lê, E. Pauwels, and S. VaiterMathematical Programming, 2025DOI arXiv
From Shortcut to Induction Head: How Data Diversity Shapes Algorithm Selection in TransformersR. Kawata, Y. Song, A. Bietti, N. Nishikawa, T. Suzuki, S. Vaiter, and D. WuAdvances in Neural Information Processing Systems (NeurIPS), 2025arXiv
Learning Theory for Kernel Bilevel OptimizationF. E. Khoury, E. Pauwels, S. Vaiter, and M. ArbelAdvances in Neural Information Processing Systems (NeurIPS), 2025arXiv
Differentiable Generalized Sliced Wasserstein PlansL. Chapel, R. Tavenard, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2025arXiv
Faster Computation of Entropic Optimal Transport via Stable Low Frequency ModesR. Chhaibi, S. Gratton, and S. VaiterPreprint, 2025arXiv

2024

A theory of optimal convex regularization for low-dimensional recoveryY. Traonmilin, R. Gribonval, and S. VaiterInformation and Inference: A Journal of the IMA, 2024arXiv
Model identification and local linear convergence of coordinate descentQ. Klopfenstein, Q. Bertrand, A. Gramfort, J. Salmon, and S. VaiterOptimization Letters, 2024arXiv
Gradient Scarcity with Bilevel Optimization for Graph LearningH. Ghanem, S. Vaiter, and N. KerivenTransactions on Machine Learning Research, 2024Featured CertificationarXiv
Convergence of Message Passing Graph Neural Networks with Generic Aggregation On Large Random GraphsM. Cordonnier, N. Keriven, N. Tremblay, and S. VaiterJournal of Machine Learning Research, 2024arXiv
Provable local learning rule by expert aggregation for a Hawkes networkS. Jaffard, S. Vaiter, A. Muzy, and P. Reynaud-BouretInternational Conference on Artificial Intelligence and Statistics (AISTATS), 2024PDF arXiv
Derivatives of Stochastic Gradient Descent in parametric optimizationF. Iutzeler, E. Pauwels, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2024arXiv
A Near-Optimal Algorithm for Bilevel Empirical Risk MinimizationM. Dagréou, T. Moreau, S. Vaiter, and P. AblinInternational Conference on Artificial Intelligence and Statistics (AISTATS), 2024PDF arXiv
How to compute Hessian-vector products?M. Dagréou, P. Ablin, S. Vaiter, and T. MoreauInternational Conference on Learning Representations (ICLR) Blogposts Track, 2024

2023

The Derivatives of Sinkhorn-Knopp ConvergeE. Pauwels, and S. VaiterSIAM Journal on Optimization, 2023DOI PDF arXiv
Supervised learning of analysis-sparsity priors with automatic differentiationH. Ghanem, J. Salmon, N. Keriven, and S. VaiterIEEE Signal Processing Letters, 2023DOI arXiv
The Geometry of Sparse Analysis RegularizationX. Dupuis, and S. VaiterSIAM Journal on Optimization, 2023DOI PDF arXiv HAL
What functions can Graph Neural Networks compute on random graphs? The role of Positional EncodingN. Keriven, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2023PDF arXiv
On the Robustness of Text VectorizersR. Catellier, S. Vaiter, and D. GarreauInternational Conference on Machine Learning (ICML), 2023arXiv
One-step differentiation of iterative algorithmsJ. Bolte, E. Pauwels, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2023SpotlightPDF arXiv
Borne inférieure de compléxité et algorithme quasi-optimal pour la minimisation de risque empirique bi-niveauxM. Dagréou, T. Moreau, S. Vaiter, and P. AblinColloque Francophone de Traitement du Signal et des Images (GRETSI), 2023
Convergence of Message Passing Graph Neural Networks with Generic Aggregation on Random GraphsM. Cordonnier, N. Keriven, N. Tremblay, and S. VaiterColloque Francophone de Traitement du Signal et des Images (GRETSI), 2023
Convergence of Message Passing Graph Neural Networks with Generic Aggregation on Random GraphsM. Cordonnier, N. Keriven, N. Tremblay, and S. VaiterGraph Signal Processing workshop, 2023

2022

Sparse and Smooth: improved guarantees for Spectral Clustering in the Dynamic Stochastic Block ModelN. Keriven, and S. VaiterElectronic Journal of Statistics, 2022DOI PDF arXiv HAL
Implicit differentiation for fast hyperparameter selection in non-smooth convex learningQ. Bertrand, Q. Klopfenstein, M. Massias, M. Blondel, S. Vaiter, A. Gramfort, and J. SalmonJournal of Machine Learning Research, 2022PDF arXiv
Benchopt: Reproducible, efficient and collaborative optimization benchmarksT. Moreau, M. Massias, A. Gramfort, P. Ablin, P.-A. Bannier, B. Charlier, M. Dagréou, T. D. l. Tour, G. Durif, C. F. Dantas, Q. Klopfenstein, J. Larsson, E. Lai, T. Lefort, M. Malezieu, B. Moufad, B. T. Nguyen, A. Rakotomamonjy, Z. Ramzie, J. Salmon, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2022PDF arXiv
A framework for bilevel optimization that enables stochastic and global variance reduction algorithmsM. Dagréou, P. Ablin, S. Vaiter, and T. MoreauAdvances in Neural Information Processing Systems (NeurIPS), 2022OralarXiv
Automatic differentiation of nonsmooth iterative algorithmsJ. Bolte, E. Pauwels, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2022PDF arXiv
Algorithmes stochastiques et réduction de variance grâce à un nouveau cadre pour l’optimisation bi-niveauxM. Dagréou, P. Ablin, S. Vaiter, and T. MoreauColloque Francophone de Traitement du Signal et des Images (GRETSI), 2022

2021

Automated data-driven selection of the hyperparameters for Total-Variation based texture segmentationB. Pascal, S. Vaiter, N. Pustelnik, and P. AbryJournal of Mathematical Imaging and Vision, 2021PDF arXiv
Linear Support Vector Regression with Linear ConstraintsQ. Klopfenstein, and S. VaiterMachine Learning, 2021DOI arXiv HAL
Block based refitting in ℓ₁₂ sparse regularisationC.-A. Deledalle, N. Papadakis, J. Salmon, and S. VaiterJournal of Mathematical Imaging and Vision, 2021DOI PDF arXiv HAL
On the Universality of Graph Neural Networks on Large Random GraphsN. Keriven, A. Bietti, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2021arXiv
From optimization to algorithmic differentiation: a graph detourS. VaiterHDR Thesis, 2021PDF

2020

Dual Extrapolation for Sparse Generalized Linear ModelsM. Massias, S. Vaiter, A. Gramfort, and J. SalmonJournal of Machine Learning Research, 2020PDF arXiv HAL
Convergence and Stability of Graph Convolutional Networks on Large Random GraphsN. Keriven, A. Bietti, and S. VaiterAdvances in Neural Information Processing Systems (NeurIPS), 2020SpotlightPDF arXiv
Implicit differentiation of Lasso-type models for hyperparameter optimizationQ. Bertrand, Q. Klopfenstein, M. Blondel, S. Vaiter, A. Gramfort, and J. SalmonInternational Conference on Machine Learning (ICML), 2020PDF arXiv

2019

Maximal Solutions of Sparse Analysis RegularizationA. Barbara, A. Jourani, and S. VaiterJournal of Optimization Theory and Applications, 2019DOI PDF arXiv HAL
Refitting solutions promoted by ℓ₁₂ sparse analysis regularization with block penaltiesC.-A. Deledalle, N. Papadakis, J. Salmon, and S. VaiterInternational Conference on Scale Space and Variational Methods in Computer Vision (SSVM), 2019arXiv HAL
Exploiting regularity in sparse Generalized Linear ModelsM. Massias, S. Vaiter, A. Gramfort, and J. SalmonSignal Processing with Adaptive Sparse Structured Representations (SPARS), 2019HAL

2018

Model Consistency of Partly Smooth RegularizersS. Vaiter, G. Peyré, and J. FadiliIEEE Transactions on Information Theory, 2018DOI PDF arXiv HAL
Optimality of 1-norm regularization among weighted 1-norms for sparse recovery: a case study on how to find optimal regularizationsY. Traonmilin, and S. VaiterNew Computational Methods for Inverse Problems (NCMIP), 2018PDF arXiv HAL
Is the 1-norm the best convex sparse regularization?Y. Traonmilin, S. Vaiter, and R. GribonvaliTWIST, 2018PDF arXiv HAL

2017

The Degrees of Freedom of Partly Smooth RegularizersS. Vaiter, C.-A. Deledalle, G. Peyré, J. Fadili, and C. DossalAnnals of the Institute of Statistical Mathematics, 2017DOI PDF arXiv HAL
CLEAR: Covariant LEAst-square Re-fitting with applications to image restorationC.-A. Deledalle, N. Papadakis, J. Salmon, and S. VaiterSIAM Journal on Imaging Sciences, 2017DOI PDF arXiv HAL
Accelerated Alternating Descent Methods for Dykstra-like problemsA. Chambolle, P. Tan, and S. VaiterJournal of Mathematical Imaging and Vision, 2017DOI PDF HAL
A Sharp Oracle Inequality for Graph-SlopeP. Bellec, J. Salmon, and S. VaiterElectronic Journal of Statistics, 2017DOI PDF arXiv HAL
Characterizing the maximum parameter of the total-variation denoising through the pseudo-inverse of the divergenceC.-A. Deledalle, N. Papadakis, J. Salmon, and S. VaiterSignal Processing with Adaptive Sparse Structured Representations (SPARS), 2017PDF arXiv HAL

2015

Model Selection with Low Complexity PriorsS. Vaiter, M. Golbabaee, J. Fadili, and G. PeyréInformation and Inference: A Journal of the IMA, 2015DOI PDF arXiv HAL
Low Complexity Regularization of Linear Inverse ProblemsS. Vaiter, G. Peyré, and J. FadiliBook chapter in Sampling Theory, a Renaissance, 2015DOI PDF arXiv HAL

2014

Stein Unbiased GrAdient estimator of the Risk (SUGAR) for multiple parameter selectionC.-A. Deledalle, S. Vaiter, J. Fadili, and G. PeyréSIAM Journal on Imaging Sciences, 2014DOI PDF arXiv HAL
Low Complexity Regularizations of Inverse ProblemsS. VaiterPhD Thesis, 2014PDF SITE

2013

Robust Sparse Analysis RegularizationS. Vaiter, G. Peyré, C. Dossal, and J. FadiliIEEE Transactions on Information Theory, 2013DOI PDF arXiv HAL
Local Behavior of Sparse Analysis Regularization: Applications to Risk EstimationS. Vaiter, C.-A. Deledalle, G. Peyré, C. Dossal, and J. FadiliApplied and Computational Harmonic Analysis, 2013DOI PDF arXiv HAL
Robust Polyhedral RegularizationS. Vaiter, G. Peyré, and J. FadiliSAMPTA, 2013PDF arXiv HAL
Stable Recovery with Analysis Decomposable PriorsJ. Fadili, G. Peyré, S. Vaiter, C.-A. Deledalle, and J. SalmonSAMPTA, 2013PDF arXiv
Robustesse au bruit des régularisations polyhédralesS. Vaiter, G. Peyré, and J. FadiliColloque Francophone de Traitement du Signal et des Images (GRETSI), 2013
Reconstruction Stable par Régularisation Décomposable AnalyseJ. Fadili, G. Peyré, S. Vaiter, C.-A. Deledalle, and J. SalmonColloque Francophone de Traitement du Signal et des Images (GRETSI), 2013
The degrees of freedom of the group Lasso for a general designS. Vaiter, G. Peyré, J. Fadili, C.-A. Deledalle, and C. DossalSignal Processing with Adaptive Sparse Structured Representations (SPARS), 2013PDF HAL

2012

Unbiased Risk Estimation for Sparse Analysis RegularizationC.-A. Deledalle, S. Vaiter, G. Peyré, J. Fadili, and C. DossalInternational Conference on Image Processing (ICIP), 2012DOI PDF HAL
Proximal Splitting Derivatives for Risk EstimationC.-A. Deledalle, S. Vaiter, G. Peyré, J. Fadili, and C. DossalNew Computational Methods for Inverse Problems (NCMIP), 2012DOI PDF HAL
The Degrees of Freedom of the Group LassoS. Vaiter, C.-A. Deledalle, G. Peyré, J. Fadili, and C. DossalInternational Conference on Machine Learning (ICML) Workshop Sparsity, Dictionaries and Projections in Machine Learning and Signal Processing, 2012PDF arXiv HAL
Risk estimation for matrix recovery with spectral regularizationC.-A. Deledalle, S. Vaiter, G. Peyré, J. Fadili, and C. DossalInternational Conference on Machine Learning (ICML) Workshop Sparsity, Dictionaries and Projections in Machine Learning and Signal Processing, 2012PDF arXiv HAL